Review of Economics and Statistics MS #15015

Title: �Does Online Availability Increase Citations? 
Theory and Evidence from a Panel of Economics and Business Journals�

Authors: Mark McCabe and Christopher Snyder

DATA/PROGRAMS/CODES for Data Archive

README file 

This file provides an outline of the programs and data provided in the
data archive that the authors used to produce the tables and figures in 
the paper.  All programs were run in Stata MP version 12.

***********************************************************************
Table 1

Program: 
	McCabe_Snyder_ReStat_Table_1.DO

Dataset: 
	McCabe_Snyder_ReStat_combined.dta

***********************************************************************
Table 2

Programs: 
	To clean data: McCabe_Snyder_ReStat_Table_2_Data_Clean.DO
	To run regressions: McCabe_Snyder_ReStat_Table_2.DO

Datasets: 
	Raw files combined in cleaning: TITLES.DTA, SUBFIELD.DTA, FaulknerReferences.xls
	Clean file used in regressions: McCabe_Snyder_ReStat_References.dta

***********************************************************************
Table 3

Programs: 
	McCabe_Snyder_ReStat_Table_3.DO

Datasets: 
	McCabe_Snyder_ReStat_combined.dta


Notes:
	Regression result is an incidence rate ratio. As indicated in the
	notes to the published paper, we manually subtract 1 from this to 
	obtain marginal effect reported in paper.

***********************************************************************
Table 4

Programs: 
	McCabe_Snyder_ReStat_Table_4.DO

Datasets: 
	McCabe_Snyder_ReStat_combined.dta

Notes:
	Regression result is an incidence rate ratio (IRR).  We convert this
	into a subscription elasticity in two steps.  First, we compute
	a marginal effect by subtracting the IRR from 1.  Then we multiply
	by mean number of subscribers. The relevant means are provided 
	by a command issued after each regression. 

***********************************************************************
Table 5

Programs: 
	To clean data: McCabe_Snyder_ReStat_Table_5_Data_Clean.DO
	To run regressions: McCabe_Snyder_ReStat_Table_5.DO

Datasets: 
	Raw files combined in cleaning: FaulknerArchiving.xls
	Clean file used in regressions: McCabe_Snyder_ReStat_archiving.dta

***********************************************************************
Table 6

Programs: 
	McCabe_Snyder_ReStat_Table_6.DO

Datasets: 
	McCabe_Snyder_ReStat_combined.dta

Notes:
	Regression result is an incidence rate ratio (IRR).  We convert this
	into a subscription elasticity in two steps.  First, we compute
	a marginal effect by subtracting the IRR from 1.  Then we multiply
	by mean number of subscribers. The relevant means are provided 
	by a command issued after each regression. 

***********************************************************************
Table 7

Programs: 
	McCabe_Snyder_ReStat_Table_7.DO

Dataset: 
	McCabe_Snyder_ReStat_combined.dta

Notes:
	Regression result is an incidence rate ratio (IRR).  We convert this
	into a subscription elasticity in two steps.  First, we compute
	a marginal effect by subtracting the IRR from 1.  Then we multiply
	by mean number of subscribers. The relevant means are provided 
	by a command issued after each regression. 

***********************************************************************
Table 8

Programs: 
	McCabe_Snyder_ReStat_Table_8.DO

Datasets: 
	McCabe_Snyder_ReStat_combined.dta

Notes:
	Reported results computed in three steps. First, the marginal effect
	is computed using the MFX command issued after each regression. This 
	figure needs to be scaled by two factors to obtain an elasticity.
	As with the elasticities in the previous tables, it needs to be 
	multiplied by subsample mean of subscribers (not the RHS variable, 
	which is subscribers x online access dummy). This is generated by 
	summ command issued after MFX. The last step is new to this table,
	required here with fractional data. Need also to divide by mean of 
	dependent variable (proportion of volume's articles cited). This is 
	reported in the header of the output from the MFX command issued 
	after the regression.

***********************************************************************
Figures

Program: 
	McCabe_Snyder_ReStat_Table_1.DO

Dataset: 
	McCabe_Snyder_ReStat_combined.dta

Notes:
	Figures 4 and 5 are schematic. Rest (1, 2, 3, 6, 7) use Stata analysis.
	This is provided in the program listed above using the listed dataset.
	Program provides documentation on computations for each figure.